From 3ba59005b92026ba846711afabaf692b87adcd81 Mon Sep 17 00:00:00 2001
From: Emiliano Vavassori <syntaxerrormmm@gmail.com>
Date: Sat, 19 Mar 2022 00:50:07 +0100
Subject: [PATCH] Versione funzionante.

---
 getversions.py | 118 ++++++++++++++++++++++++++++++-------------------
 1 file changed, 72 insertions(+), 46 deletions(-)

diff --git a/getversions.py b/getversions.py
index be6eaf3..cf09a20 100644
--- a/getversions.py
+++ b/getversions.py
@@ -3,60 +3,86 @@
 
 import urllib.request
 from lxml import etree
-import re, sys
+from packaging.version import parse as parse_version
+import re, sys, json
 
-archiveurl = "https://downloadarchive.documentfoundation.org/libreoffice/old/"
-baseurl = "https://download.documentfoundation.org/libreoffice/"
+ARCHIVE = "https://downloadarchive.documentfoundation.org/libreoffice/old/"
+RELEASE = "https://download.documentfoundation.org/libreoffice/stable/"
+DAILY = "https://dev-builds.libreoffice.org/daily/master/Linux-rpm_deb-x86_64@tb87-TDF/current/"
 
 def getlatestrel(basever):
-    """Search in downloadarchive the latest version matching the baseversion indicated."""
-    versionlist = etree.HTML(urllib.request.urlopen(archiveurl).read()).xpath('//td/a')
+    """Search in downloadarchive for the latest version matching baseversion."""
+    versionlist = etree.HTML(urllib.request.urlopen(ARCHIVE).read()).xpath('//td/a')
     # Getting a more polished matching list
-    cleanlist = list(dict.fromkeys([x.strip('/') if x.startswith(basever) else None for x in versionlist]))
-    # Removing None entries
-    cleanlist.remove(None)
+    cleanlist = list(dict.fromkeys([x.text.strip('/') for x in versionlist if x.text.startswith(basever)]))
+
     # Sorting, then returning the last version
     return sorted(cleanlist)[-1]
 
-if len(sys.argv) > 1:
-    # A version has been specified.
-    version = sys.argv[1]
-    if '.' in version:
-        # Numbered version.
-        # Let's check it is a 4 dotted release
-        if not len(version.split('.')) == 4:
-            version = getlatestrel(version)
+def getbranchrel(branch):
+    """Based on branch names, get the release number."""
+    basedirurl = {}
+    version = ''
+    if branch == 'daily':
+        basedirurl = { 'x86_64': DAILY, 'x86': '-' }
+        version = etree.HTML(urllib.request.urlopen(DAILY).read()).xpath('//td/a')[1].text.split('_')[1]
+
+        return { 'version': version, 'basedirurl': basedirurl }
+        
+    versions = etree.HTML(urllib.request.urlopen(RELEASE).read()).xpath('//td/a')
+    index = 1
+    if branch ==  'still':
+        index = 2
+    elif branch == 'fresh':
+        index = 3
+    version = getlatestrel(versions[index].text.strip('/'))
+
+    return { 'version': version, 'basedirurl': getbaseurl(version) }
+
+def getbaseurl(version):
+    """Returns the links based on the numeric version."""
+    basediriurl = {}
+    url = ARCHIVE + '/' + version + '/deb/'
+    # x86 binaries are not anymore offered after 6.3.0.
+    if parse_version(version) < parse_version('6.3.0'):
+        basedirurl['x86'] = url + 'x86/'
     else:
-        # string version
-        if version == 'still':
-        elif version == 'fresh':
-        elif version == 'daily':
+        basedirurl['x86'] = '-'
+    
+    basedirurl['x86_64'] = url + 'x86_64/'
 
+    return basedirurl
 
-# Running for stable channel
-html_stable = urllib.request.urlopen(baseurl + 'stable/').read()
-
-# Processing page content
-stable_contents = etree.HTML(html_stable)
-
-# Processing first page links
-stable_versions = []
-for link in stable_contents.xpath('//td/a'):
-    next if link.text == "Parent Directory"
-
-    stable_versions.append(link.get('href').replace('/', '')
-
-    # getting URL for specific versions
-    for arch in [ 'x86', 'x86_64' ]:
-        newurl = baseurl + 'stable/' + link.get('href') + 'deb/' + arch + '/'
-        res = etree.HTML(urllib.request.urlopen(newurl).read()).xpath("//td/a")
-        if len(res) == 1:
-            # No packages provided
-            next
-
-        for link in res:
-            next if res.text == 'Parent Directory'
-            next if re.search(r'deb.tar.gz$', res.text):
-                # Matches a package - good.
-                # Checking for language packs and offline help
+if __name__ == '__main__':
+    # Preparing variables for outputting
+    version = ''
+    basedirurl = {}
+    basever = 'fresh'
+    
+    # At the end of the checks, we need a version string and a basedirurl, which
+    # should be a dictionaly for x86, x86_64 with the base of the directory where
+    # to download the files.
+    if len(sys.argv) > 1:
+        # A version has been specified.
+        basever = sys.argv[1]
+    
+    # Once overridden with Argv, parse the inputs
+    if '.' in basever:
+        # Numbered version.  Let's check it is a 4 dotted release
+        if len(basever.split('.')) == 4:
+            version = basever
+        else:
+            version = getlatestrel(basever)
+        
+        basedirurl = getbaseurl(version)
+    else:
+        # String versions.
+        a = getbranchrel(basever)
+        version = a['version']
+        basedirurl = a['basedirurl']
 
+    output = """query: %s
+version: %s
+x86: %s
+x86_64: %s"""  % (basever, version, basedirurl['x86'] or '-', basedirurl['x86_64'])
+print(output)